Sure! Here’s the American English translation of the provided text:
San José (California), March 18, 2025 — NVIDIA announced at GTC the expansion of its Omniverse platform, consolidating it as an operating system for physical AI, backed by major industrial software and service providers such as Ansys, Databricks, Dematic, Omron, SAP, Schneider Electric, ETAP, and Siemens.
This expansion aims to accelerate large-scale industrial digitalization through advanced simulation, massive generation of synthetic data, and integration of AI models for robotics and industrial automation.
New Blueprints for Robot-Ready Factories
The company has launched new Omniverse Blueprints, connected to the foundational models of the physical world through NVIDIA Cosmos™, allowing users to create virtual environments to test fleets of robots and generate synthetic data at scale.
One such blueprint, Mega, which is already available in preview, allows for the simulation of operations with multiple robots in industrial digital twins. Companies like Schaeffler, Hyundai Motor Group, Mercedes-Benz, Pegatron, and Foxconn are already using Mega to test and train humanoid robots and industrial manipulators in simulated environments before their actual deployment.
For example, Foxconn is simulating fleets of robots for its factories of the future, while KION Group, Dematic, and idealworks are integrating Mega to optimize advanced warehouse automation and robotic fleet management.
Digital Twins for AI Factories
NVIDIA has also introduced a blueprint for designing and simulating digital twins of AI factories, where engineers can optimize the layout of servers, cooling, and energy consumption. Companies like Cadence and Schneider Electric with ETAP have integrated their simulation tools into this new framework, and suppliers such as Vertiv have made available simulation-ready (SimReady) 3D models of their power and climate control systems.
Moreover, the Isaac GR00T Blueprint, designed to generate synthetic data for training humanoid robots, is now available for developers, drastically reducing data collection times and accelerating advancements in advanced robotics.
An Operating System for Physical AI Breaking Industrial Silos
The traditional industrial ecosystem suffers from data compartmentalization. NVIDIA Omniverse, based on the OpenUSD standard, facilitates the integration of physical and digital data, enabling developers to work on a common platform. Companies like Ansys, Cadence, Hexagon, Omron, Rockwell Automation, and Siemens are already incorporating Omniverse’s interoperability and visualization capabilities into their simulation and automation solutions.
For its part, General Motors has begun using Omniverse to optimize its industrial processes and train robotic platforms for operations such as welding and material handling. At the end of the production chain, Unilever is adopting physically accurate digital twins to streamline the creation of advertising content.
Omniverse in the Cloud for Everyone
To facilitate the development and deployment of applications based on OpenUSD, NVIDIA has announced the availability of Omniverse as pre-configured virtual images on AWS Marketplace and Microsoft Azure Marketplace. It will soon also be available on Oracle Cloud Infrastructure and Google Cloud, using the recently introduced NVIDIA RTX PRO 6000 Blackwell Server Edition.
Towards a Common Language in Robotics
In collaboration with Disney Research and Intrinsic, NVIDIA has introduced the OpenUSD Asset Structure Pipeline for Robotics, a data structure that unifies robotic workflows, facilitating the transition from digital twins to deployment in physical hardware.
The expansion of Omniverse reinforces NVIDIA’s commitment to being the central axis of industrial digitalization and the development of physical AI, supporting major industry leaders in creating smart factories and collaborative environments driven by advanced artificial intelligence.
Source: Nvidia